A method for optimizing finish rolling thickness control of martensitic stainless steel

Abstract

The application belongs to the technical field of industrial manufacturing, and provides a kind of martensitic stainless steel finishing thickness control optimization method, including data collection and arrangement, the data collected and arranged are cleaned and preprocessed, variables are screened using L1 regularization algorithm, a regression model is established and an optimization scheme is proposed by using the regression model combined with the process conditions in production;The application screens out the key variables affecting the thickness change through L1 regularization, and optimizes the parameter setting in the finishing process by using the regression model, which can significantly reduce the phenomenon of thin head and thick tail of the strip steel, thereby improving the thickness consistency and precision of the final product, avoiding the product scrap rate caused by non-standard thickness, thereby improving the production efficiency and reducing the production cost.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of industrial manufacturing, and relates to optimization of a finish rolling control method, in particular to a martensitic stainless steel finish rolling thickness control optimization method. BACKGROUND

[0002] Martensitic stainless steel is widely used in various industrial manufacturing due to its unique physical properties and structure, especially in fields requiring high strength and corrosion resistance materials such as chemical industry, nuclear power, aerospace, etc. During the production and manufacturing process of martensitic stainless steel, especially in the finish rolling stage, an important challenge is how to accurately control its thickness, especially when the thickness is less than 3.0mm, the thickness control of the head and tail of the finished strip often deviates, causing the finished product thickness to be out of standard, affecting subsequent processing and use, such as Figure 1 As shown in the actual thickness curve of the strip, the thickness deviates significantly over time.

[0003] The specific phenomenon is that the head thickness is too thin and the tail thickness is too thick. This phenomenon is not only related to temperature, pressure, and plastic deformation of the material, but also affected by the response speed, gain setting, and other factors of the AGC (automatic thickness control system) in the finish rolling process.

[0004] In order to solve these problems, traditional thickness control methods such as increasing AGC gain and manually adjusting tail compensation coefficient have little effect. SUMMARY

[0005] The purpose of the present application is to provide a martensitic stainless steel finish rolling thickness control optimization method, based on L1 regularization algorithm, to construct a linear regression model and establish a quantitative relationship between variables, so as to screen out the variable with the greatest impact on thickness change, optimize the control parameters of each link in the finish rolling process, and ensure that the final product thickness meets the target requirements.

[0006] The technical solution adopted by the present application to achieve the above purpose is:

[0007] A martensitic stainless steel finish rolling thickness control optimization method, comprising the following steps:

[0008] S1, data collection and arrangement, including:

[0009] Pressure changes at different stages of the rolling process; by monitoring the load pressure data, the influence of different pressures on the finish rolling thickness of stainless steel is analyzed;

[0010] AGC output data; collect AGC total output, absolute value AGC output, relative AGC output and monitoring AGC output data, and analyze the actual influence of the data on thickness control;

[0011] Temperature data; collect real-time data of stainless steel finishing mill inlet and outlet temperature;

[0012] Thickness gauge data; relationship between outlet thickness data of thickness gauge and other variables.

[0013] S2, cleaning and preprocessing the collected data, deleting noise data and abnormal values.

[0014] S3, screening variables by using L1 regularization algorithm, by introducing a penalty term, the coefficients of unimportant variables become zero, so as to screen out variables that have significant influence on the result, the specific steps are as follows:

[0015] S31, establishing a preliminary model: through a multiple linear regression model, a preliminary relationship between thickness and multiple variables is established;

[0016] S32, introducing L1 regularization term: introducing L1 regularization term in the regression model, controlling the complexity of the model, reducing the influence of multicollinearity;

[0017] S33, variable screening: through model training, screening out variables with non-zero coefficients, these variables are the key variables that have greater influence on thickness change.

[0018] S4, establishing a regression mode; after screening out the key variables, a linear regression model is constructed to determine the quantitative relationship between each variable and the thickness of the stainless steel finishing mill, and the specific steps are as follows:

[0019] S41, data division: the collected data is divided into training set and test set to ensure the generalization ability of the model;

[0020] S42, model training: using the data in the training set, the linear regression model is trained to determine the regression coefficients of each variable;

[0021] S43, model verification: using the test set data to verify the model, evaluating the prediction accuracy and reliability of the model;

[0022] S44, result analysis: by analyzing the size and sign of the regression coefficients, the positive and negative influence of different variables on thickness change is judged.

[0023] S5, using the regression model to propose an optimization scheme combined with the process conditions in production.

[0024] The method has the following beneficial effects:

[0025] 1、The method of the present application can significantly reduce the phenomenon of thin head and thick tail of the strip steel by screening the key variables affecting the thickness change through L1 regularization and optimizing the parameter setting in the finishing rolling process by using the regression model, thereby improving the thickness consistency and precision of the final finished stainless steel product, avoiding the product scrap rate caused by the non-standard thickness, and improving the production efficiency and reducing the production cost;

[0026] 2、The method of the present application improves the response speed and gain setting of the finishing rolling AGC system based on the data analysis and modeling results, so that the finishing rolling system can respond to the thickness change faster and more accurately, and reduce the thickness deviation in the production process;

[0027] 3、The method of the present application introduces data analysis and machine learning technology, so that the finishing rolling process parameter setting is more automated and intelligent, and the dependence on manual intervention and experience adjustment is reduced. BRIEF DESCRIPTION OF DRAWINGS

[0028] Figure 1 is a flow chart of the method of the present application. DETAILED DESCRIPTION

[0029] The method of the present application will be further described below in combination with the drawings and examples:

[0030] The method of the present application is applied on a hot continuous rolling 2250 finishing rolling mill, and the finishing rolling thickness control of the martensitic stainless steel is optimized based on L1 regularization and regression model in the big data background.

[0031] The optimization method comprises the following steps:

[0032] S1, data collection and arrangement: in the finishing rolling process of stainless steel, a large number of variables will affect the final thickness of the finishing rolling stainless steel strip, in order to comprehensively analyze these variables, a large amount of production variable data needs to be collected first, including:

[0033] 1) the pressure change at different stages in the rolling process, specifically the pressure feedback of F1 to F7; by monitoring the load pressure data, the influence of different pressures on the finishing rolling thickness of stainless steel is analyzed. The pressure change at different stages in the rolling process has a direct influence on the plastic deformation of the material, and is an important variable affecting the thickness;

[0034] 2) AGC output data; the automatic adjustment function of the AGC system plays a key role in thickness control. The data of AGC total output, absolute value AGC output, relative AGC output and monitoring AGC output are collected to analyze the actual influence of the data on the thickness control;

[0035] 3) Temperature data; Collect real-time data of stainless steel finishing rolling inlet temperature and outlet temperature; Temperature is an important factor affecting the deformation behavior of materials. In the finishing rolling process, the change of inlet temperature and outlet temperature will directly affect the thickness of the strip steel;

[0036] 4) Thickness gauge data; Relationship between outlet thickness data of thickness gauge and other variables; By analyzing the relationship between the outlet thickness data of F7 thickness gauge as the response variable and other variables, the main factors affecting the thickness change can be found.

[0037] S2, After the data collection is completed, clean and pretreat the collected and sorted data, delete noise data and outliers, and ensure the accuracy and reliability of the data.

[0038] S3, Use L1 regularization algorithm to screen variables, introduce penalty term to make unimportant variable coefficients zero, and screen out variables that have significant impact on results, the specific steps are as follows:

[0039] S31, Establish a preliminary model: establish a preliminary relationship between thickness and multiple variables through a multiple linear regression model;

[0040] S32, Introduce L1 regularization term: introduce L1 regularization term in the regression model to control the complexity of the model and reduce the impact of multicollinearity;

[0041] S33, Variable screening: through model training, screen out variables with non-zero coefficients, these variables are the key variables that have greater impact on thickness change;

[0042] The variables screened by L1 regularization can significantly simplify the model structure, while improving the interpretability and predictive ability of the model.

[0043] S4, Establish a regression model;

[0044] After screening the key variables, construct a linear regression model to determine the quantitative relationship between each variable and the thickness of stainless steel finishing rolling, the specific steps are as follows:

[0045] S41, Data division: divide the collected data into training set and test set to ensure the generalization ability of the model;

[0046] S42, Model training: use the data in the training set to train the linear regression model and determine the regression coefficients of each variable;

[0047] S43, Model verification: use the test set data to verify the model and evaluate the prediction accuracy and reliability of the model;

[0048] S44, Result analysis: analyze the size and sign of the regression coefficients to determine the positive and negative impact of different variables on thickness change.

[0049] S5. Based on the actual production process conditions, propose optimization solutions.

[0050] By establishing a regression model, we can identify which factors have a significant impact on the problem of a thinner head and a thicker tail, thus providing a basis for adjusting process parameters.

[0051] Based on the analysis results of the regression model, the parameter settings of the finishing mill AGC system are optimized. Specific optimization suggestions are proposed for AGC response speed, gain settings, and tail compensation coefficient:

[0052] Increase AGC response speed; by improving the response speed of the AGC system, the frequency of thickness deviations can be reduced.

[0053] Adjust the tail compensation coefficient; to address the issue of excessive tail thickness, optimize the tail compensation coefficient to ensure that the thickness remains within the target range during the tail rolling process;

[0054] Enhance AGC gain settings; based on traditional gain settings, further increase the gain value of the AGC system to ensure that the system can respond quickly and effectively to thickness changes.

[0055] Based on data analysis and modeling results, the above steps improved the response speed and gain settings of the AGC system, enabling the system to respond to thickness changes faster and more accurately, and reducing thickness deviations during the production process.

[0056] After the model and parameters were optimized, the effectiveness of the optimization scheme was verified through field tests. By applying the optimized AGC parameter settings on the actual production line, the significant improvement in the phenomenon of thinner head and thicker tail was verified. The consistency and accuracy of the final thickness control were evaluated by comparing and analyzing the data from the thickness gauge.

[0057] The method of this invention uses L1 regularization to screen out key variables affecting thickness variation and uses regression models to optimize parameter settings in the finishing rolling process. This can significantly reduce the phenomenon of strip being too thin at the head and too thick at the tail, thereby improving the thickness consistency and accuracy of the final product, avoiding product scrap rate due to substandard thickness, thus improving production efficiency and reducing production costs.

[0058] This invention, by introducing data analysis and machine learning techniques, automates and intelligently sets parameters during the finishing rolling process, reducing reliance on manual intervention and experience-based adjustments. This method is not only applicable to the finishing rolling process of martensitic stainless steel, but can also be extended to the thickness control and optimization of other types of stainless steel or metallic materials, demonstrating broad industrial application prospects.

Claims

1. A method of optimizing finish rolling gauge control of a martensitic stainless steel, characterized by: Comprising the following steps: S1, data collection and arrangement, comprising: 1) pressure changes at different stages of the rolling process; by monitoring load pressure data, the effects of different pressures on stainless steel finishing thickness are analyzed; 2) AGC output data; collect AGC total output, absolute AGC output, relative AGC output and monitor AGC output data, analyze the actual impact of data on thickness control; 3) temperature data; collect real-time data of stainless steel finishing inlet temperature and outlet temperature; 4) thickness gauge data; the relationship between the outlet thickness data of the thickness gauge and other variables; S2, clean and pretreat the collected and arranged data, delete noise data and abnormal values; S3, use L1 regularization algorithm to screen variables, by introducing a penalty term, the coefficients of unimportant variables become zero, thereby screening out variables that have a significant impact on the result; S4, establish a regression model; after screening the key variables, a linear regression model is constructed to determine the quantitative relationship between each variable and the stainless steel finishing thickness; S5, use the regression model combined with the process conditions in production to propose an optimization scheme.

2. A method of finishing thickness control optimization of a martensitic stainless steel according to claim 1, characterized in that: The specific steps of S3 include: S31, establish a preliminary model: establish a preliminary relationship between thickness and multiple variables through a multiple linear regression model; S32, introduce L1 regularization term: introduce L1 regularization term in the regression model to control the complexity of the model and reduce the impact of multicollinearity; S33, variable screening: through model training, screen out variables with non-zero coefficients, which are the key variables that have a greater impact on thickness change.

3. The method for finishing thickness control optimization of a martensitic stainless steel according to claim 1, characterized in that: The specific steps of S4 include: S41, data division: divide the collected data into training set and test set to ensure the generalization ability of the model; S42, model training: use the data in the training set to train the linear regression model to determine the regression coefficients of each variable; S43, model verification: use the test set data to verify the model and evaluate the prediction accuracy and reliability of the model; S44, result analysis: by analyzing the size and sign of the regression coefficients, the positive and negative effects of different variables on thickness change are determined.

Images