An online camber control method based on joint drive of mechanism data

Abstract

The present invention discloses an online camber control method based on a mechanism-data-driven approach. The method comprises: collecting characteristic process parameters related to the camber of a hot-rolled intermediate billet on-site; calculating characteristic parameters based on the mechanism of the rolling plastic deformation process; processing the obtained characteristic parameter data to obtain a sample data set; establishing a mechanism-data-driven regression prediction model for the camber of a hot-rolled intermediate billet based on the sample data set; using the regression prediction model to predict the camber value of the slab in the final rolling pass in real time, obtaining the minimum predicted camber value corresponding to the operator's leveling value, and using the roll gap leveling value corresponding to the minimum predicted camber value as the optimal roll gap leveling value for the current pass. Roller tilting and leveling are then performed based on the predicted minimum camber value, thereby achieving automatic control of the camber of the hot-rolled intermediate billet. The present invention can effectively reduce manual intervention and achieve automatic control of the camber of the hot-rolled intermediate billet.

Description

Technical Field

[0001] The present invention relates to the technical field of intermediate billet camber control in the rough rolling stage of plate and strip rolling, and in particular to an online camber control method based on mechanism data joint drive. Background Art

[0002] With the rapid development of my country's iron and steel metallurgical technology, the demand for high-precision, high-tech and high-value-added steel is also increasing. As the mainstream research direction in the field of modern strip rolling, the control of asymmetric plate shape has always attracted much attention. Due to the asymmetric extension of the slab in the width direction during the rolling process, the slab bends to one side and forms a sickle bend. The asymmetric sickle bend of the slab poses a great obstacle to the control and stability of the finishing precision, and may even cause steel piling accidents, causing serious damage. Therefore, studying the influencing factors and control strategies for the formation of sickle bend in hot-rolled slabs is of great significance for improving the quality of hot-rolled plate shape.

[0003] With the development of flatness control theory, two mainstream methods have emerged for controlling asymmetric flatness. One method indirectly measures the curvature of the slab by measuring the pressure difference between the left and right sides, and then controls the slab by adjusting the roll gap. The other method involves installing a detection device to detect the amount of slab curvature at the inlet, and then analyzing and adjusting the rolls to control the slab. Numerous complex on-site factors complicate mathematical modeling for precise control of the intermediate slab's camber. The intermediate slab is subject to numerous nonlinear factors during the rough rolling process, and existing camber mechanism control models struggle to accurately describe these nonlinear characteristics. Traditional and conventional camber control methods struggle to meet the camber control requirements of hot rolling. Summary of the Invention

[0004] The present invention provides an online camber control method based on joint drive of mechanism data to solve the technical problems that the existing camber mechanism control model is difficult to accurately describe the nonlinear characteristics and the traditional and conventional camber control method is difficult to meet the camber control requirements of hot rolling site.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] In one aspect, the present invention provides a camber online control method based on a joint drive of mechanism data, the camber online control method based on a joint drive of mechanism data comprising:

[0007] On-site collection of characteristic process parameters related to camber of hot-rolled intermediate billet;

[0008] Calculating characteristic parameters based on the rolling plastic deformation process mechanism based on the characteristic process parameters;

[0009] Processing the calculated characteristic parameter data to obtain a sample data set;

[0010] Based on the sample data set, a hot rolling intermediate slab camber regression prediction model based on the mechanism data joint drive is established; wherein the input of the regression prediction model is the characteristic parameters based on the rolling plastic deformation process mechanism, and the output is the prediction result of the slab camber value of the target rolling pass;

[0011] Real-time prediction of the camber value of the slab in the final rolling pass is performed using the regression prediction model, and a minimum camber prediction value corresponding to the operator leveling value is obtained based on the output of the regression prediction model;

[0012] Obtain the roll gap leveling value corresponding to the minimum value of the sickle bend prediction value, use the roll gap leveling value corresponding to the minimum value of the sickle bend prediction value as the optimal roll gap leveling value for this pass, send the optimal roll gap leveling value to the rough rolling basic automation control system, execute roll tilting and leveling, and realize automatic control of the sickle bend of the hot rolling intermediate billet.

[0013] Furthermore, the hot rolling intermediate billet sickle camber related characteristic process parameters include: geometric variables, process variables, and material and equipment variables; wherein,

[0014] The geometric variables include: slab bending value at the exit of pass 1, slab bending value at the exit of pass 3, hydraulic cylinder width, slab width, slab thickness at the exit of pass 1, slab thickness at the exit of pass 2, and slab thickness at the exit of pass 3;

[0015] The process variables include: 1-pass roll leveling value, 2-pass roll leveling value, 3-pass roll leveling value, 1-pass rolling mill drive side and operating side rolling forces, and 2-pass rolling mill drive side and operating side rolling forces;

[0016] The material and equipment variables include: the longitudinal stiffness of the three-pass rolling mill and the plastic deformation coefficient of the three-pass rolling slab.

[0017] Furthermore, the calculation is based on characteristic parameters of the rolling plastic deformation process mechanism, including:

[0018] Determining characteristic process parameters related to camber of the hot rolling intermediate billet that affect the target variable; wherein the target variable is the camber value of the slab at the three-pass exit; the characteristic process parameters related to camber of the hot rolling intermediate billet that affect the target variable include: the camber value of the slab at the one-pass exit, the hydraulic cylinder width and the slab width, the thickness of the slab at the first three-pass exit, the leveling value of the rolling rolls in the first three passes, the rolling forces on the operating side and the transmission side of the rolling mill in the one-pass and two-pass rolling, the longitudinal stiffness of the rolling mill in the three-pass rolling, and the plastic deformation coefficient of the slab in the three-pass rolling;

[0019] The characteristic parameters based on the rolling plastic deformation process mechanism are calculated using the following formula:

[0020]

[0021]

[0022]

[0023]

[0024]

[0025] Among them, C1 and C3 are the bending values ​​of the slab at the exit of 1 pass and 3 passes respectively, W is the slab width, L is the length of the hydraulic cylinder, H1 and H3 are the thickness of the slab at the exit of 1 pass and 3 passes respectively, ΔS1, ΔS2, ΔS3 are the leveling values ​​of the rolling rollers in the first 3 passes, F OS1 ,F DS1 ,F OS2 ,F DS2 are the rolling forces on the operating side and transmission side of the rolling mill for 1 and 2 passes respectively, K is the longitudinal stiffness of the rolling mill for 3 passes, and Π1~Π9 are characteristic parameters determined based on the mechanism of the rolling plastic deformation process.

[0026] Furthermore, the calculated characteristic parameter data is processed to obtain a sample data set, including:

[0027] The experimental sample data set is obtained by calculation based on the processed characteristic parameters;

[0028] The obtained experimental sample data are spliced ​​and sorted according to the time series, and the abnormal data values ​​of the spliced ​​and sorted experimental sample data are eliminated according to the 3σ criterion to obtain a sample data set.

[0029] Furthermore, based on the sample data set, a hot rolling intermediate billet camber regression prediction model based on the joint drive of mechanism data is established, including:

[0030] Dividing the sample data set to obtain input data and output data;

[0031] A regression model is constructed, and the constructed regression model is trained based on the input data and the output data to obtain a hot rolling intermediate billet camber regression prediction model driven jointly by the mechanism data.

[0032] Furthermore, the sample data set is divided to obtain input data and output data, including:

[0033] Set the sliding window width according to the timing;

[0034] For the sample data set, the input data of different combinations and the output corresponding to the next data of the input data are divided according to the set sliding window width as output data for model training.

[0035] Furthermore, the sliding window width ranges from 50 to 500.

[0036] Furthermore, a regression model is constructed and trained based on the input data and output data to obtain a hot rolling intermediate billet camber regression prediction model driven by the mechanism data, including:

[0037] Load the calculated input data and output data, and perform normalization on the loaded input data and output data to obtain a normalized data set; wherein the normalization interval is [-1, 1];

[0038] For a given data set S = {(x i ,y i )|i=1,2,…,n},x i ∈R n is the n-dimensional input feature parameter, y i ∈R is the output target parameter, which is mapped by nonlinear Mapping the original data x to a high-dimensional feature space, the regression function is as follows:

[0039]

[0040] Where w and b are weight vector and bias respectively, is a nonlinear mapping function;

[0041] Given the penalty parameter C>0 and the insensitive loss function ε>0, the regression model is expressed as:

[0042]

[0043]

[0044] Where, ξ, * is a slack variable, C is a penalty parameter, ε is an insensitive loss function, y i is the output target parameter, ξ i , is the i-th slack variable;

[0045] Calculate the regression model and introduce the kernel function κ(x c ,x i ):

[0046]

[0047] Where δ is the kernel function parameter, x c is the center of the kernel function, x i is the input feature parameter;

[0048] The final regression model is as follows:

[0049]

[0050] Where, is the support vector;

[0051] The finally obtained regression model is trained based on the input data and the output data to obtain a hot rolling intermediate billet camber regression prediction model driven jointly by the mechanism data.

[0052] Furthermore, the resulting regression model is trained based on the input data and output data, including:

[0053] Initialize the model parameters;

[0054] Iteratively optimize the penalty parameters and kernel function parameters of the model;

[0055] In the iterative optimization process, the objective function is calculated, that is, the sum of squares of the differences between the model prediction value and the true value is minimized, and the model performance evaluation index is calculated;

[0056] When the preset conditions are met, the iteration is stopped and the model parameter optimization results are obtained.

[0057] Furthermore, the roll gap leveling value corresponding to the minimum camber prediction value is obtained, and the roll gap leveling value corresponding to the minimum camber prediction value is used as the optimal roll gap leveling value for this pass, specifically:

[0058] The roll gap leveling value corresponding to the minimum predicted value of the slab sickle bend in the last pass is obtained by a traversal search method, and the roll gap leveling value corresponding to the minimum predicted value of the sickle bend is used as the optimal roll gap leveling value for this pass.

[0059] On the other hand, the present invention further provides an electronic device, comprising a processor and a memory; wherein the memory stores at least one instruction, and the instruction is loaded and executed by the processor to implement the above method.

[0060] In yet another aspect, the present invention further provides a computer-readable storage medium, wherein the storage medium stores at least one instruction, and the instruction is loaded and executed by a processor to implement the above method.

[0061] The beneficial effects brought about by the technical solution provided by the present invention include at least:

[0062] 1. The present invention can reduce labor intensity: The present invention provides the optimal roll leveling value in a data-driven manner, reduces the occurrence of camber and operator intervention, and realizes automatic control of camber of hot-rolled intermediate billets;

[0063] 2. The present invention can improve the control effect: The present invention provides an online control algorithm for sickle camber based on the joint drive of mechanism data, which establishes an accurate prediction and online control method for the sickle camber of the hot-rolled intermediate billet, and can more accurately predict the sickle camber and perform online control while reducing labor. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0065] Figure 1 1 is a schematic diagram of the execution flow of the sickle camber online control method based on the joint drive of mechanism data provided by an embodiment of the present invention;

[0066] Figure 2 It is a schematic diagram of dividing the experimental data provided by an embodiment of the present invention into sliding windows of different widths. DETAILED DESCRIPTION

[0067] To make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0068] First embodiment

[0069] This embodiment provides a sickle bend online control method based on the joint drive of mechanism data, which can be implemented by electronic equipment. Figure 1 As shown, the following steps are included:

[0070] S1, on-site collection of hot rolling intermediate billet sickle related characteristic process parameters;

[0071] Specifically, in this embodiment, the collected characteristic process parameters related to the sickle bending of the hot-rolled intermediate billet include: geometric variables, process variables, and material and equipment variables; wherein, the geometric variables include: the bending value of the slab at the outlet of 1 pass, the bending value of the slab at the outlet of 3 passes, the width of the hydraulic cylinder, the width of the slab, the thickness of the slab at the outlet of 1 pass, the thickness of the slab at the outlet of 2 passes, and the thickness of the slab at the outlet of 3 passes; the process variables include: the leveling value of the roller at 1 pass, the leveling value of the roller at 2 passes, the leveling value of the roller at 3 passes, the rolling force of the transmission side and the operating side of the rolling mill at 1 pass, and the rolling force of the transmission side and the operating side of the rolling mill at 2 passes; the material and equipment variables include: the longitudinal stiffness of the rolling mill at 3 passes and the plastic deformation coefficient of the slab rolled at 3 passes.

[0072] S2, calculating characteristic parameters based on the rolling plastic deformation process mechanism based on the characteristic process parameters;

[0073] Specifically, in this embodiment, the implementation process of the above S2 is as follows:

[0074] S21, determining the hot rolling intermediate billet sickle related characteristic process parameters that affect the target variable; wherein, the target variable is the slab bending value of the three-pass outlet; the hot rolling intermediate billet sickle related characteristic process parameters that affect the target variable are specifically: C1, D, W, H1, H2, H3, ΔS1, ΔS2, ΔS3, F OS1 ,F DS1 ,F OS2 ,F DS2 ,K,Q; where C1 is the bending value of the slab at the exit of the first pass, in mm; D and W are the width of the hydraulic cylinder and the slab width, in mm; H1, H2, H3 are the thickness of the slab at the exit of the first three passes, in mm; ΔS1, ΔS2, ΔS3 are the leveling values ​​of the rolling rollers in the first three passes, in mm; F OS1 ,F DS1 ,F OS2 ,F DS2 are the rolling forces on the operating side (OS) and the driving side (DS) of the rolling mill for 1 and 2 passes, respectively, in kN; K and Q are the longitudinal stiffness of the rolling mill for 3 passes and the plastic deformation coefficient of the slab for 3 passes, respectively, in kN / mm;

[0075] S22, calculate the characteristic parameters based on the rolling plastic deformation process mechanism, the calculation formula is as follows:

[0076]

[0077]

[0078]

[0079]

[0080]

[0081] Among them, C1 and C3 are the bending values ​​of the slab at the exit of the 1st and 3rd passes respectively, in mm, D and W are the width of the hydraulic cylinder and the slab respectively, in mm, L is the length of the hydraulic cylinder, in mm, H1, H2, H3 are the thickness of the slab at the exit of the first 3 passes, in mm, ΔS1, ΔS2, ΔS3 are the leveling values ​​of the rolling rollers in the first 3 passes, in mm, F OS1 ,F DS1 ,F OS2 ,F DS2 are the rolling forces on the operating side (OS) and the drive side (DS) of the rolling mill for 1 and 2 passes, respectively, in kN. K and Q are the longitudinal stiffness of the rolling mill for 3 passes and the plastic deformation coefficient of the slab for 3 passes, respectively, in kN / mm. Π1 to Π9 are characteristic parameters determined based on the mechanism of the rolling plastic deformation process.

[0082] S3, processing the calculated characteristic parameter data to obtain a sample data set;

[0083] Specifically, in this embodiment, the implementation process of the above S3 is as follows:

[0084] S31, calculating and obtaining an experimental sample data set according to the processed characteristic parameters;

[0085] S32, splicing and sorting the obtained experimental sample data according to the time series, and eliminating abnormal data values ​​from the spliced ​​and sorted experimental sample data according to the 3σ criterion to obtain a sample data set.

[0086] Specifically, this embodiment splices and sorts 3558 experimental sample data in time series, and removes outliers according to the 3σ criterion to obtain 3500 experimental data.

[0087] S4, establishing a hot rolling intermediate slab camber regression prediction model based on the sample data set and jointly driven by the mechanism data; wherein the input of the regression prediction model is a characteristic parameter based on the rolling plastic deformation process mechanism, and the output is a prediction result of the slab camber value of the target rolling pass;

[0088] Specifically, in this embodiment, the implementation process of the above S4 is as follows:

[0089] S41, dividing the sample data set to obtain input data and output data;

[0090] Specifically, if Figure 2As shown, this embodiment, for a sample data set, forms a sliding window of a certain width into input data Input_data and output data Output_data according to time sequence; it includes:

[0091] S411, setting the sliding window width sw according to the time sequence, where sw ranges from 50 to 500;

[0092] Specifically, in this embodiment, assuming that sw=250, Input_data is shifted with a width of 250 to obtain different input data sets, and the corresponding next (251) output data Output_data is used for model testing.

[0093] S412: For the sample data set, the input data of different combinations and the output corresponding to the next data of the input data are divided according to the set sliding window width as output data for model training.

[0094] S42, constructing a regression model, and training the constructed regression model based on the input data and the output data to obtain a hot rolling intermediate billet camber regression prediction model based on the joint drive of the mechanism data, which includes:

[0095] S421, load the calculated input data and output data, and perform normalization processing on the loaded input data and output data; wherein, the normalization interval is [-1, 1], and the normalization mapping used is as follows:

[0096]

[0097] Where x′ is the normalized variable data, x is the original variable data, x_min = min(x), x_max = max(x);

[0098] S422, for a given data set S={(x i ,y i )|i=1,2,…,n},x i ∈R n is the n-dimensional input feature parameter, y i ∈R is the output target parameter, which is mapped by nonlinear Mapping the original data x to a high-dimensional feature space, the regression function is as follows:

[0099]

[0100] Where w and b are weight vector and bias respectively, is a nonlinear mapping function;

[0101] S423, given the penalty parameter C>0 and the insensitive loss function ε>0, the regression model is expressed as:

[0102]

[0103]

[0104] Where, ξ, * is a slack variable, C is a penalty parameter, ε is an insensitive loss function, y i is the output target parameter, ξ i , is the i-th slack variable;

[0105] S424, calculate the regression model and introduce the kernel function κ(x c ,x i ):

[0106]

[0107] Where δ is the kernel function parameter, x c is the center of the kernel function, x i is the input feature parameter;

[0108] S425, the final regression model f(x) is as follows:

[0109]

[0110] Where, is the support vector;

[0111] S426, training the finally obtained regression model based on the input data and the output data to obtain a hot rolling intermediate billet camber regression prediction model based on the joint drive of the mechanism data, which includes:

[0112] S4261, initializing model parameters;

[0113] S4262, iteratively optimizing the model parameters (C and δ);

[0114] S4263, during the iterative optimization process, calculate the objective function, i.e., minimize the sum of squares of the differences between the model prediction value and the true value, and calculate the model performance evaluation index;

[0115] S4264: When the preset conditions are met, the iteration is stopped and the model parameter optimization result is obtained.

[0116] S5, predicting the camber value of the slab in the final rolling pass in real time using the regression prediction model, and obtaining a minimum camber prediction value corresponding to the operator leveling value based on the output of the regression prediction model;

[0117] S6, obtain the roll gap leveling value corresponding to the minimum camber prediction value, use the roll gap leveling value corresponding to the minimum camber prediction value as the optimal roll gap leveling value for this pass, send the optimal roll gap leveling value to the rough rolling basic automation control system, execute roll tilting and leveling, and realize automatic control of the camber of the hot rolling intermediate billet.

[0118] Specifically, in this embodiment, the above steps are specifically as follows: the sickle bend of the last rolling pass slab is accurately predicted in real time through the established sickle bend regression prediction model, and the minimum sickle bend prediction value corresponding to the operator's leveling value is obtained according to the model output; then the roll gap leveling value corresponding to the minimum sickle bend value of the last slab is obtained through the traversal search method as the optimal roll gap leveling value of this pass, and it is sent to the rough rolling basic automation control system to execute the roll tilting and leveling to realize the automatic control of the sickle bend of the hot rolling intermediate slab.

[0119] In summary, this embodiment provides an online camber control method based on a joint drive of mechanism data. This method analyzes and processes the multivariable and nonlinear characteristic parameters affecting the camber of hot-rolled slabs based on the mechanism of the rolling plastic deformation process. This reduces the variable dimension while also shortening the runtime of the machine learning model. A traversal search method is then used to determine the roll gap leveling value corresponding to the minimum camber value of the slab in the final pass. This value is then transmitted to the roughing basic automation control system to execute roll tilting and leveling. This achieves the goal of online control of the camber quality of hot-rolled intermediate slabs based on a joint drive of mechanism data, effectively reducing manual intervention.

[0120] Second embodiment

[0121] This embodiment provides an electronic device, which includes a processor and a memory; wherein the memory stores at least one instruction, and the instruction is loaded and executed by the processor to implement the method of the first embodiment.

[0122] The electronic device may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) and one or more memories, wherein the memory stores at least one instruction, which is loaded by the processor to execute the above method.

[0123] Third embodiment

[0124] This embodiment provides a computer-readable storage medium storing at least one instruction, which is loaded and executed by a processor to implement the method of the first embodiment described above. The computer-readable storage medium may be a ROM, random access memory, CD-ROM, magnetic tape, floppy disk, or optical data storage device. The instructions stored therein can be loaded by a processor in a terminal to execute the method described above.

[0125] Furthermore, it should be noted that the present invention may be provided as a method, apparatus, or computer program product. Thus, embodiments of the present invention may take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of the present invention may take the form of a computer program product embodied on one or more computer-usable storage media containing computer-usable program code.

[0126] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, an embedded processor, or other programmable data processing terminal device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the process in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0127] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device, so that a series of operation steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for implementing the process in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0128] It should also be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or terminal device comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or terminal device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or terminal device comprising the element.

[0129] Finally, it should be noted that the above is a preferred embodiment of the present invention. It should be noted that although the preferred embodiment of the present invention has been described, it is clear that those skilled in the art, once they understand the basic inventive concept of the present invention, can make various improvements and modifications without departing from the principles of the present invention. Such improvements and modifications should also be considered as within the scope of protection of the present invention. Therefore, the appended claims are intended to be interpreted as including the preferred embodiment and all changes and modifications that fall within the scope of the embodiments of the present invention.

Claims

1. A camber online control method based on mechanism data joint drive, characterized in that: include: On-site collection of characteristic process parameters related to camber of hot-rolled intermediate billet; Calculating characteristic parameters based on the rolling plastic deformation process mechanism based on the characteristic process parameters; Processing the calculated characteristic parameter data to obtain a sample data set; Based on the sample data set, a hot rolling intermediate slab camber regression prediction model based on the mechanism data joint drive is established; wherein the input of the regression prediction model is the characteristic parameters based on the rolling plastic deformation process mechanism, and the output is the prediction result of the slab camber value of the target rolling pass; Real-time prediction of the camber value of the slab in the final rolling pass is performed using the regression prediction model, and a minimum camber prediction value corresponding to the roll gap leveling value is obtained based on the output of the regression prediction model; Obtain the roll gap leveling value corresponding to the minimum value of the sickle bend prediction value, use the roll gap leveling value corresponding to the minimum value of the sickle bend prediction value as the optimal roll gap leveling value for this pass, send the optimal roll gap leveling value to the rough rolling basic automation control system, execute roll tilting and leveling, and realize automatic control of the sickle bend of the hot rolling intermediate billet.

2. The camber online control method based on mechanism data joint drive according to claim 1 is characterized in that: The hot rolling intermediate billet sickle related characteristic process parameters include: geometric variables, process variables, and material and equipment variables; wherein, The geometric variables include: slab bending value at the exit of pass 1, slab bending value at the exit of pass 3, hydraulic cylinder width, slab width, slab thickness at the exit of pass 1, slab thickness at the exit of pass 2, and slab thickness at the exit of pass 3; The process variables include: 1-pass roll leveling value, 2-pass roll leveling value, 3-pass roll leveling value, 1-pass rolling mill drive side and operating side rolling forces, and 2-pass rolling mill drive side and operating side rolling forces; The material and equipment variables include: the longitudinal stiffness of the three-pass rolling mill and the plastic deformation coefficient of the three-pass rolling slab.

3. The camber online control method based on mechanism data joint drive according to claim 2 is characterized in that: The calculation is based on characteristic parameters of the rolling plastic deformation process mechanism, including: Determining characteristic process parameters related to camber of the hot rolling intermediate billet that affect the target variable; wherein the target variable is the camber value of the slab at the three-pass exit; the characteristic process parameters related to camber of the hot rolling intermediate billet that affect the target variable include: the camber value of the slab at the one-pass exit, the hydraulic cylinder width and the slab width, the thickness of the slab at the first three-pass exit, the leveling value of the rolling rolls in the first three passes, the rolling forces on the operating side and the transmission side of the rolling mill in the one-pass and two-pass rolling, the longitudinal stiffness of the rolling mill in the three-pass rolling, and the plastic deformation coefficient of the slab in the three-pass rolling; The characteristic parameters based on the rolling plastic deformation process mechanism are calculated using the following formula: Among them, C1 and C3 are the bending values ​​of the slab at the exit of 1 pass and 3 passes respectively, W is the slab width, L is the length of the hydraulic cylinder, H1 and H3 are the thickness of the slab at the exit of 1 pass and 3 passes respectively, ΔS1, ΔS2, ΔS3 are the leveling values ​​of the rolling rollers in the first 3 passes, F OS1 ,F DS1 ,F OS2 ,F DS2 are the rolling forces on the operating side and transmission side of the rolling mill for 1 and 2 passes respectively, K is the longitudinal stiffness of the rolling mill for 3 passes, and Π1~Π9 are characteristic parameters determined based on the mechanism of the rolling plastic deformation process.

4. The camber online control method based on mechanism data joint drive according to claim 1 is characterized in that: The step of processing the calculated characteristic parameter data to obtain a sample data set includes: The experimental sample data set is obtained by calculation based on the processed characteristic parameters; The obtained experimental sample data are spliced ​​and sorted according to the time series, and the abnormal data values ​​of the spliced ​​and sorted experimental sample data are eliminated according to the 3σ criterion to obtain a sample data set.

5. The camber online control method based on mechanism data joint drive according to claim 1 is characterized in that: Based on the sample data set, a hot rolling intermediate billet camber regression prediction model based on joint driving of mechanism data is established, including: Dividing the sample data set to obtain input data and output data; A regression model is constructed, and the constructed regression model is trained based on the input data and the output data to obtain a hot rolling intermediate billet camber regression prediction model driven jointly by the mechanism data.

6. The camber online control method based on mechanism data joint drive according to claim 5 is characterized in that: Dividing the sample data set to obtain input data and output data includes: Set the sliding window width according to the timing; For the sample data set, the input data of different combinations and the output corresponding to the next data of the input data are divided according to the set sliding window width as output data for model training.

7. The camber online control method based on mechanism data joint drive according to claim 6, characterized in that: The sliding window width ranges from 50 to 500.

8. The camber online control method based on mechanism data joint drive according to claim 5, characterized in that: A regression model is constructed, and the constructed regression model is trained based on the input data and the output data to obtain a hot rolling intermediate billet camber regression prediction model driven by the mechanism data, including: Load the calculated input data and output data, and perform normalization on the loaded input data and output data to obtain a normalized data set; wherein the normalization interval is [-1, 1]; For a given data set S = {(x i ,y i )|i=1,2,…,n},x i ∈R n is the n-dimensional input feature parameter, y i ∈R is the output target parameter, which is mapped by nonlinear Mapping the original data x to a high-dimensional feature space, the regression function is as follows: Where w and b are weight vector and bias respectively, is a nonlinear mapping function; Given the penalty parameter C>0 and the insensitive loss function ε>0, the regression model is expressed as: Where, ξξ * is a slack variable, C is a penalty parameter, ε is an insensitive loss function, y i is the output target parameter, ξ i , is the i-th slack variable; Calculate the regression model and introduce the kernel function κ(x c ,x i ): Where δ is the kernel function parameter, x c is the center of the kernel function, x i is the input feature parameter; The final regression model is as follows: Where, is the support vector; The finally obtained regression model is trained based on the input data and the output data to obtain a hot rolling intermediate billet camber regression prediction model driven jointly by the mechanism data.

9. The camber online control method based on mechanism data joint drive according to claim 8, characterized in that: Training the resulting regression model based on the input data and the output data includes: Initialize the model parameters; Iteratively optimize the penalty parameters and kernel function parameters of the model; In the iterative optimization process, the objective function is calculated, that is, the sum of squares of the differences between the model prediction value and the true value is minimized, and the model performance evaluation index is calculated; When the preset conditions are met, the iteration is stopped and the model parameter optimization results are obtained.

10. The camber online control method based on mechanism data joint drive according to claim 1, characterized in that: The roller gap leveling value corresponding to the minimum camber prediction value is obtained, and the roller gap leveling value corresponding to the minimum camber prediction value is used as the optimal roller gap leveling value for this pass, specifically: The roll gap leveling value corresponding to the minimum predicted value of the slab sickle bend in the last pass is obtained by a traversal search method, and the roll gap leveling value corresponding to the minimum predicted value of the sickle bend is used as the optimal roll gap leveling value for this pass.

Images